Pentera Opens Early Access to Agentic Web App and API Pentesting

Unite.AI
Read full story· 5 min
Pentera has launched early access to Pentera Apps, an agentic web app and API penetration-testing tool, after beta tests with over 60 customers revealed critical security flaws. The tool automates testing of application behaviors, workflows, and APIs, supporting various authentication methods and protocols. General availability is planned for Q1 2027, following a July 2026 beta announcement targeting Q4 2026.

What happened

On November 18, 2026, Pentera announced early access to Pentera Apps, a new tool for automated penetration testing of web applications and APIs. Over 60 customers participated in the beta phase, which uncovered vulnerabilities such as unpaid e-commerce checkouts, account takeovers, and data extraction through unauthenticated APIs. The tool is designed to test application behaviors and workflows using AI agents that simulate attacks. General availability is now scheduled for the first quarter of 2027, after a broader beta phase planned for late 2026.

Key details

  • Beta testing involved more than 60 customers.
  • Critical issues found included multifactor-authentication bypass and SQL injection in APIs.
  • The tool supports REST, GraphQL, and WebSocket interfaces, as well as various authentication methods.
  • A live demonstration is scheduled for December 1, 2026, at 11:00 a.m. ET.
  • Customer data is processed in assigned EU or US regions, with zero retention and no training on customer data.

Why it matters

Security and application teams at organizations using web apps and APIs can now automate penetration testing between manual assessments, potentially reducing undetected vulnerabilities as software changes. The tool provides concrete exploit evidence and supports continuous validation, but does not replace manual testing or code scanners. Customers must manage the findings and remediation process, with the option to retest after fixes, while maintaining control over test scope and data handling.


More in Cybersecurity